Audio Vs MIDI recording - newbie question

So I’m relatively new to iOS recording and one thing that confuses me is the fact that most software instruments can record as both MIDI and audio. I fully understand the difference between the two, and the editing possibilities of MIDI vs audio, but the last time I was recording regularly the lines between MIDI and Audio were much clearer: if the sound was generated by the computer/soundcard it was MIDI, anything else was audio.

Besides the different editing possibilities and the opportunity to sample audio, what are the main reasons for recording an AU instrument as MIDI or audio?
Is there much difference in CPU use?

Comments

  • There are some differences for CPU and storage...
    If you record as MIDI then when playing back the CPU is needed by the AU to generate the audio from the MIDI, the synth engine will need to be running to do this.
    If you record as audio then the CPU is only required for the actual audio playback, most of which is needed whether playing back MIDI or Audio, so in general terms audio playback is lighter on CPU than MIDI playback....
    However, Audio recording and playback requires a lot more information to be fed into it than MIDI does, so there is a lot more writing/reading of that information during record/playback, and it uses a lot more space on the device....to some extent this is swings and roundabouts.

    There are less obvious benefits to audio recording as well....if you have an AU recorded as Audio, you no longer need to keep or find the preset you used within the AU, which is important particularly if you are collaborating and sharing stuff with someone else....as MIDI they would need the same AU and same preset, with audio they would need neither.

    One thing to remember, you can easily turn MIDI into Audio......turning audio into MIDI is not a simple task...

    The best compromise is having the ability to 'freeze' your MIDI tracks, like in Cubasis, this means you can keep the MIDI for editing, but convert to audio when needed for performance....but then easily go back to the MIDI for either further editing, or to claim back space on your device...

  • Great explanation!

    Question: if I freeze a track in Cubasis will the loaded instance of the AU synth still suck cpu power?

    @AndyPlankton said:
    There are some differences for CPU and storage...
    If you record as MIDI then when playing back the CPU is needed by the AU to generate the audio from the MIDI, the synth engine will need to be running to do this.
    If you record as audio then the CPU is only required for the actual audio playback, most of which is needed whether playing back MIDI or Audio, so in general terms audio playback is lighter on CPU than MIDI playback....
    However, Audio recording and playback requires a lot more information to be fed into it than MIDI does, so there is a lot more writing/reading of that information during record/playback, and it uses a lot more space on the device....to some extent this is swings and roundabouts.

    There are less obvious benefits to audio recording as well....if you have an AU recorded as Audio, you no longer need to keep or find the preset you used within the AU, which is important particularly if you are collaborating and sharing stuff with someone else....as MIDI they would need the same AU and same preset, with audio they would need neither.

    One thing to remember, you can easily turn MIDI into Audio......turning audio into MIDI is not a simple task...

    The best compromise is having the ability to 'freeze' your MIDI tracks, like in Cubasis, this means you can keep the MIDI for editing, but convert to audio when needed for performance....but then easily go back to the MIDI for either further editing, or to claim back space on your device...

  • @Tones4Christ said:
    Great explanation!

    Question: if I freeze a track in Cubasis will the loaded instance of the AU synth still suck cpu power?

    @AndyPlankton said:
    There are some differences for CPU and storage...
    If you record as MIDI then when playing back the CPU is needed by the AU to generate the audio from the MIDI, the synth engine will need to be running to do this.
    If you record as audio then the CPU is only required for the actual audio playback, most of which is needed whether playing back MIDI or Audio, so in general terms audio playback is lighter on CPU than MIDI playback....
    However, Audio recording and playback requires a lot more information to be fed into it than MIDI does, so there is a lot more writing/reading of that information during record/playback, and it uses a lot more space on the device....to some extent this is swings and roundabouts.

    There are less obvious benefits to audio recording as well....if you have an AU recorded as Audio, you no longer need to keep or find the preset you used within the AU, which is important particularly if you are collaborating and sharing stuff with someone else....as MIDI they would need the same AU and same preset, with audio they would need neither.

    One thing to remember, you can easily turn MIDI into Audio......turning audio into MIDI is not a simple task...

    The best compromise is having the ability to 'freeze' your MIDI tracks, like in Cubasis, this means you can keep the MIDI for editing, but convert to audio when needed for performance....but then easily go back to the MIDI for either further editing, or to claim back space on your device...

    No. That is the point of track freezing. While a track is frozen (true in Cubasis and Auria Pro), the audio is used and the plug-in isn't. If the "plug-in" is actually an IAA app, you may want to eject it from memory if it is one that sucks CPU when idle.

  • Sweet! I always ejected the AU Synth thinking it was using CPU. Saves me time in settings and preset setup! Thanks 👍🏼

  • I think it is always worth pointing out that there isn't a "right" way for things like this. It really is "what works best for you" (& obviously you can only make that choice in full knowledge of the implications as described above).

    I know some people prefer to get things into audio as soon as possible since they are happier working with audio - other people keep it as midi for as long as possible since they might want to edit it at some point. But it's all workflow and your choice should be "what suits your way of working".

    I suspect the lines are now less clear for this very reason - give people more choices.

  • I like midi better as well but with limited cpu I and this amazing new apps requiring more cpu juice there’s no choice but to freeze up and turn to audio, good thing is that Cubasis allows you to unfreeze and just copy the produced audio file into another track in case you changed your mind about keeping it; but also allows you to go back and do more midi edits to the original! I really like that.

  • wimwim
    edited June 14

    Besides the different editing possibilities and the opportunity to sample audio, what are the main reasons for recording an AU instrument as MIDI or audio?
    Is there much difference in CPU use?

    The simplest answer I can give is:

    • Each MIDI note processed has to be played by a synth. Playing a synth consumes CPU. Some synths require a lot of CPU processing, some less, but generally speaking, playing a synth is going to consume more CPU than playing audio. It’s like the difference between making soup from scratch vs. opening a can.
    • However, while a track is still MIDI, it’s much, much easier to to change. You can fix timing of notes, change them, copy and paste patterns and add variation, change the sound of the synth, or even the synth itself, change the key, tempo, etc. far easier.
    • MIDI takes a tiny fraction of the space to store vs. audio.
    • Audio can be re-used anywhere. So can MIDI, but then you need to set up the synths to process it.
    • MIDI can be tempo-changed without any impact to the sound. Audio tempo can be changed by some apps, but there’s always some degradation of quality when doing so.
    • Freezing is a great compromise as long as the host supports it well. The caveat with IAA apps is the settings won’t be remembered, so you need remember to save a preset before freezing and reload it if you unfreeze the track.
  • @pagefall said:
    I think it is always worth pointing out that there isn't a "right" way for things like this. It really is "what works best for you" (& obviously you can only make that choice in full knowledge of the implications as described above).

    I know some people prefer to get things into audio as soon as possible since they are happier working with audio - other people keep it as midi for as long as possible since they might want to edit it at some point. But it's all workflow and your choice should be "what suits your way of working".

    I suspect the lines are now less clear for this very reason - give people more choices.

    I tend to freeze as I go and duplicate that audio to its own track. That way I have the midi available if I need it.

    It has proven handy for projects that lay dormant for years. Even though some of the original instruments were no longer supported, I still had the audio stems.

Sign In or Register to comment.